Résumé : BACKGROUND: The purpose of this study was to carry out two randomised phase II trials of S16020, a new olivacine derivative, tested as a single agent in patients with recurrent head and neck cancer, using methotrexate as the control arm to validate the results. PATIENTS AND METHODS: S16020 at either 80 or 100 mg/m2 was administered as a 3-h infusion every 3 weeks. Methotrexate, 40 or 50 mg/m2, was given by bolus injection, weekly for a minimum of 6 weeks. In total, 36 patients were entered in the randomised studies (25 in an initial study, 11 in a confirmatory study) of whom 24 received S16020 and 12 received methotrexate. RESULTS: A scheduled interim analysis showed one patient having a non-confirmed objective response with S16020 and three patients having a confirmed objective response with methotrexate. In the methotrexate group, there were no patients with severe non-haematological toxicity. With S16020, there was a high incidence of severe non-haematological toxicities, including asthenia, oedema of the face, oedema and pain at the tumour sites and erythematous rash; consequently, both studies were stopped. CONCLUSIONS: Both studies were stopped due to the poor anticipated benefit/risk ratio for S16020, although time to progression and overall survival time were similar in both treatment arms.